Combining Range-Suffrage and Sort-Mid Algorithms for Improving Grid Scheduling

Abstract

Grid scheduling is one of the most known NP-complete problems. Heterogeneity of machines causes mapping of tasks to be a challenging problem. Several meta-heuristic algorithms have been designed to reach optimality as possible. Sort-Mid is a recent efficient scheduler of excellent resource utilization. Its strategy was based on computing the mean of two consecutive middle values in the sorted list of completion time. Thereafter, it maps the task with the maximum mean to the machine having the minimum completion time. However, Range-Suffrage scheduler obtains a better makespan. It was built on searching for the maximum average value of completion times among certain tasks. These tasks were selected according to their suffrage values under specified constraint. This paper proposes RSSM as highly efficient scheduler. It combines Sort-Mid and Range-Suffrage algorithms to achieve the maximum resource utilization and the minimum makespan. RSSM methodology is based on sorting only the completion time corresponding to tasks satisfying Range-Suffrage constraint. Experimental tests manifest the superiority of proposed algorithm over the two original meta- heuristics and other promising algorithms such as Min-Min.